Hacker News Reader

At the same time, the majority of those running Windows have the same relationship to it as you have with Linux on your DSL modem - "It's what the computer I bought comes with" - and yet we are not discounting those Windows users. Same goes for macOS, especially now where Asahi Linux is the only and yet incomplete alternative.

In some comparisons it makes sense to remove DSL modems from the equation and focusing on some more "computer-like" subset of device (otherwise no OS would ever do better than "several orders of magnitude fewer deployments than Linux"), but discounting every device where you did not actively pick the OS would make for an extremely biased comparison.

Plus, we don't care about such distinctions outside comparisons. SteamOS have already driven significant improvements for regular desktop users. Same goes for Tizen, Android and ChromeOS. Even wonky DSL firmwares have positive effects for the rest of us.

> GNU+Linux

no.

Exactly. And MacBook users don't choose MacOS either.

For many non tech savvy users if you give them a Linux laptop with any modern DE, they can't tell the difference.

But it seems these filters are almost always cherry picked against Linux for some reason I sincerely don't know.

They would, as soon as they try to use any software most people on their circle of friends use, CD/DVD coming on the hardware they bought at the mall, or any computer magazine besides Linux Format and similar.

My dad exclusively use ie(due to activex/flash in some internal corp site), excel, and word on the laptop I bought for him.

And the only thing he use it for entertainment is browser (for reading novels)

If I am managed to setup all these properly on other system. I doubt him would ever notice a difference.

Non tech user just can't care about what system he is using less. All matter for him are tools that help him to finish his work.

What about dvd/cd or install some new softwares? Well, I am the in house tech support.

That is the thing, he only manages that thanks to the in house tech support, he would be lost without it.

> My dad exclusively use ie(due to activex/flash in some internal corp site), excel, and word on the laptop I bought for him.

That was a majority use-case between 2005-2015. Now it's a niche case (yeah, even if absolute numbers are still big enough, relatively speaking it's a niche)

Well, in 2023 a majority of that would be web or electron apps so they might not be able to tell the difference. They would be equally stumped at the sight of a CD/DVD no matter their choice of OS, as machines have not shipped with optical drives in ages.

Most normal people don't buy new software all the time, most of them like to keep their old copies going as much as possible.

Most desktop computers still do, and in alternative there are USB stick variants.

Which doesn't change the fact regarding what drivers, for what set of OSes, are supported out of the box.

Desktop computers are unusual nowadays, despite being a decent option. They are gone from non-gaming retail stores from what I can tell, so it's much more of an active decision to get.

It is true that if you try to reuse your 2003 Office install you will fail, but such unsupported and deprecated software will also cause trouble on a modern PC. Even if it runs, using an old version of Outlook is extremely unsafe...

Common end-user hardware does not require drivers nowadays either (not even printers due to IPPAnywhere and co., even though manufacturers still ship them for some reason).

Things are always hairy outside that though - macOS no longer permit kernel extensions, and you know you are in a dark place if DKMS gets involved on Linux.

macOS does not permit kernel extensions, because contrary to GNU/Linux they take ABI and kernel stability seriously, and are incrementally turning the OS into a proper microkernel, with all extensions running in userspace.

Do you have a source on Apple working to turn macOS into a proper microkernel? Couldn't find anything. Certainly Mach is a microkernel, but Darwin very much isn't. I do know that Apple has worked to expose some kernel-level hardware interfacing features in safer ways than giving full kernel extension access, but otherwise that's it?

How do you call a kernel whose drivers and extensions are all in user space?

WWDC sessions on that roadmap state quite clearly that is the long term end goal, all the kernel extension mechanisms will only be available in userspace, with one year transition for each subsystem after an userspace API is made available.

Except many if not most (at least in the past) chose a Mac for the software first and the hardware second. I know I did. And I’ve also been using desktop Linux begrudgingly beside it, ever since it existed. Windows too, though I’ve lately removed any need for that.

> Exactly. And MacBook users don't choose MacOS either.

I'm not sure what would make you an authoritative source on why people by Macs or where you're sourcing your thoroughly researched data, but I can tell you I purchased an MBA to run macOS and the software that runs on macOS. Or like the Intel N100 mini PC I purchased explicitly to run proxmox + OPNSense. Or like my Windows laptop to have a mobile lab.

I pick the machine based on the software I want to run. I'm sure we can find one, perhaps if we stretch it, two other people on HN who also purchase machines based off of the OS/software they need to run.

Many on HN might be intimately familiar with macOS internals, but we are not in any way or form the average users. Running OPNSense on Proxmox definitely sets the "outlier" sticky bit.

Most users do not watch WWDC and do not know what OS release notes are. They don't know what the boundary between their web browser and their OS is, and macOS just becomes "the thing that nags them to update it" - a nag that users unfortunately still ignore, as evident by my recent confrontation with Big Sur machines.

Oh I completely agree with you that folks on HN are likely to be of a particular mindset when it comes to technology purchases.

However, that doesn't mean that every non-HN user out there buys the aluminum-shell-in-the-vague-shape-of-a-laptop and doesn't make a conscious choice.

My point still stands. The GP doesn't have the data to back up the assertion.

There is also no data to back up the opposite assertion. It would be rather surprising to see comments on HN based on hard statistical data - an implicit "IMO" prefix makes sense to apply in most human discussions.

> That does not mean that every non-HN user...

Not every - but I do believe it is still the majority. There are a lot of people out there, and the knowledge required to make the aforementioned OS choice is niche.

This is not implying stupidity, just that there are many trades and interests out there, and subscribing to ours specifically is not a given.

> There is also no data to back up the opposite assertion.

Which is why I made no assertion with a percentage.

> but I do believe it is still the majority

What makes you believe that? Thousands of data points? Hundreds? Tens? The handful of people around you? Certainly not statistics.

You are indeed making the opposite assertion by implying the statement is false, hypocritically with no information to back it up. By your standards, your counter is entirely invalid. "I do not know" is how you avoid making an assertion.

Ridiculous standards aside, I find it an extremely reasonable to assert that given more than 7 billion people across vast areas, interests, ideologies and jobs, and given the vast hi.an knowledge, expertise and culture, that any particular interest or knowledge is only shared by a small subset.

> You are indeed making the opposite assertion by implying the statement is false

That's not at all what they said. All they're saying -- correctly -- is that you are asserting some sort of magnitude ("majority") without any data to back it up.

I would guess, though, from my personal experience, that you are probably right that a majority of people just get whatever laptop with whatever OS they're used to because that's what their parents/school/employer gave them to use, and when it's time for a new machine, they just get whatever they had before. But I don't think this is a very large majority.

> I find it an extremely reasonable to assert that given more than 7 billion people [...] that any particular interest or knowledge is only shared by a small subset.

Probably true, but also remember that OS choice isn't always driven by interest. A macOS user may get frustrated with the state of gaming on the Mac, and decide to switch to Windows. Or a Windows user might really want or need to use an application only available on macOS (though I expect this sort of thing doesn't happen as often anymore, since more and more of people's computer use ends up being through a web browser). A Windows user might also buy an iPhone or iPad and get into the Apple ecosystem enough that they decide to switch to macOS.

Certainly some people who do switch OSes don't do so because they've made an independent choice; they do so because they switch employers, and something else is the only thing available, or a friend evangelizes another OS to them to the point they want to give it a try, and end up liking it.

Regardless, many people these days don't even have a laptop or desktop computer, and do all their computing on their phone or tablet. I think that 7 billion number gets a lot smaller when you consider that. (Also, as an aside, the current world population is estimated to be a bit over 8 billion now, not 7.)

The problem with making unreasonable demands for data in response to casual discussions is that it is usually done when the person strongly disagrees with the statement. For that reason, it is implicitly a counter argument, and a hypocritical one: "Your opinion differs from mine, so you must provide data to back up yours!". None of the available options can be considered default, so any outcome is equally "grand" and subject to same requirements. Opposing without having a standpoint could happen in a peer review for a paper, but that is not what this is.

Your response is more reasonable, and is also more out in the open about the alternate belief (of course equally without data). Nothing wrong with disagreeing - only about making up unbalanced burdens of proof in casual discussions.

> a nag that users unfortunately still ignore

For good reason. Running those updates requires restarting the computer, which in turn means you have to get your working state setup again after each update. People use computers to get things done, and running updates often has the opposite effect.

That and updates are going to be the time where computers stop working. If you don't change anything, it will generally continue to work fine for years. My experience with MacOS at work was that I could expect to spend at least a day fixing whatever broke after each major OS version update, so naturally I'd put off doing it for as long as possible.

Even if things continue to work, updates may come with UI changes that disrupt the user's workflow (and the trend tends to be to remove or hide functionality and make things less useful/information dense as time goes by). Unless there's something you specifically want, modern software updates tend to be high risk low reward from the user's perspective.

Incidentally this is why I prefer desktop Linux. With the exception of Firefox turning into Chrome over the years, FOSS software tends to be remarkably stable. My home computer feels like it hasn't changed at all in almost a decade. I never have to fiddle with it.

What was surprising to me was that having an Ubuntu Laptop and a Windows 11 PC side by side that Ubuntu needs about twice as often restarts compared to Windows. I only have that setup 3-4 months, so it could be an outlier though.

Does it force the updates? Ot can you postpone them indefinitely?

My biggest issue with Windows was the forced reboots.

> They don't know what the boundary between their web browser and their OS is,

Exactly. The thing many users are referring to as MacOS, is in fact, Darwin/MacOS, or as I've recently taken to calling it, Darwin plus MacOS. MacOS is not an operating system unto itself, but rather another component of a fully functioning operating system made useful by the BSD corelibs, shell utilities and vital system components comprising a full OS as defined by the Single Unix Specification v3.

I'm not sure what point you're trying to make

Would it help if I used English?

But at that point then who gives a damn what the “desktop” or other market share is because it’s essentially meaningless and can be manipulated into whatever you want it to be. Add in the fact that a lot of what you run whether SaaS or even a shell is in a server somewhere else. The actual client operating system is increasingly about as relevant for a lot of use cases as what software your terminal ran a few decades ago.

"And MacBook users don't choose MacOS either."

I don't see why. I really like macOS. More than any other environment it's uncluttered, well organised, it gets out of the way. It abstracts a lot of unnecessary technical detail but still allows diving into that, if I'm inclined. And most of all, I like well integrated system-wide features instead of palming it off to developers. Sure the hardware is good too, but I don't know where you get that idea from.

Eh, I'd say two things:

It's VERY EXTREMELY IMPORTANT to "teach the controversy" of GNU+Linux. It's not at all important to actually use the term.

It can be quite pragmatic to use it too, there's no need to be religious about it. Even though the term gets criticized for "what about other parts than GNU", "what about Alpine" etc. it's actually a very useful term which makes people know exactly what you mean behind it, which is not something that can be said about "Linux" without providing additional context.

Exactly. We got GNU/Linux, we got GNU/kWindows, we got DalvikJavaWhatever/Linux etc etc

well, maybe you are right and it's not as clear cut. But I think people very intentionally buy a PC or a Mac which to most means the hardware and the operating system together. And, not to put too fine a point on it, but NOBODY here is counting mac user as BSD users either.

I agree with this assessment but the same two senses exist for Iphones and Macs too. To some extent it is even the same with Windows. A lot of people just buy _a_ laptop. They don't even know what an operating system is.

They know to the extent of the applications they need to use, when accessing goverment sites, and if it doesn't work it is broken, from their point of view.

Having said this, the local library from my German city does run on GNU/Linux.

I would be very curious to see the numbers comparing intentional OS installation. How many people go out of their way to buy and install Windows (thereby truly choosing Windows) versus how many people go out of their way to download and install Linux.

If Linux wins in that category as well, could we finally stop moving the goalposts?

You know, I used to be a Windows guy but Windows really jumped the shark some years ago and so I am just preparing to move over to Linux at the next best time. This is just to say, I want to the best for Linux and I don't want to move the goalposts unfairly. But I do think there is diffence between "using" Linux as in Android and using Linux as in Arch and I think the people who use Windows use it more similarly to that later way.

> or are just using whats one the phone they wanted

In that case 99.9% of windows users fall in the same category. They are just using the OS that came on the device they bought. How many people actually bought a separate license, and downloaded a Windows installer because they specifically wanted to install Windows on their device?

How many people actually bought a separate license, and downloaded a Windows installer because they specifically wanted to install Windows on their device?

Pretty much the entire (DIY) PC gaming market? Granted they may not buy Windows but they are certainly making the choice to download and use it over linux.

Sure there is a pre-built market for desktop PCs but I would think the PC Gaming market skews to DIY though I have no stats.

Note: I run arch btw.

Yeah I know. But they must be less than 0.5% of the total number of Windows users. Probably even less. So you are just proving my point here that practically all windows users just use the OS that comes with the device they got.

The story you are commenting on is about a hardware survey of gamers. I think quite a bit more than 0.5% chose to use Windows intentionally.

No, that used to be the case a few years ago. Proton runs practically any game on Linux these days. No good reason to pick windows these days

proton is good, but it's not nearly that good.

Yes, it doesn't run 5% of games smoothly therefore it's shite. Give me a break

30% actually.

Just a few weeks ago, I tried playing a Japanese game that I was interested in, but it didn't even get past the start menu. Not a big deal for me as I'm not an avid gamer, I just moved on.

But imagine telling a hardcore gamer that no, you can't play that new game released an hour ago. Maybe it'll become compatible a few months later, maybe not.

Or tell a professional artist that no, they can't use that art software anymore, they have to retrain their entire skillset with an entirely different software.

Now that's a show-stopper. For most people, it's 100% compatibility or bust.

The community is really responsive, such that you could point the hardcore gamer to the forums and they would likely get help getting it fixed.

Professional artists are also an interesting choice. If they are in a major art studio, they are probably having to use custom tooling there already. And Wacom is very well supported on linux.

It's responsive but most people aren't so technically skilled.

I have a game that works fine with proton on one machine, but doesn't on another machine. But works fine with windows on both machines.

I tried for hours to get it to work on the other machine, since it clearly can. No luck.

In the end with proprietary software, it's always a black box.

They aren't necessarily technically skilled, but they are often building the skills. Is literally how a lot of us got into computers. Did we actually understand what himem was? Not at all, but we could play with DOS and friends really well. :D

This is also an odd rabbit hole to fight about. I've had so many windows boxes that couldn't play games throughout the years that I've basically accepted that games programming is hard. :D

Except that Windows is the target OS emulated by Proton, which otherwise wouldn't have any games to run.

Yes, and what's your point?

It is a failure regarding fostering the development of Linux games.

No difference of being a nice package for something like MAME.

Your whole comment is hilarously incorrect because overwhelming majority of windows users never intentionally installed it. It just came pre-installed with their new laptop. The exact scenario you invoke with a DSL device.

Do you have stats on that? I would imagine if you swapped it out with linux or BSD they would have something to say about why all their apps aren't available. It's a choice.

A modem's application is the one thing it does and the interface is an abstraction layer standardised enough that the modem could be replaced in a shoe closet without anyone even noticing.